Dynamic analysis tools are often used by computer software developers to test computer software applications, typically by first exploring a computer software application to discover its interfaces, including those by which data may be provided to the application, and then by interacting with the application's interfaces and monitoring the application's responses to such interactions. In one type of dynamic analysis a computer software application is tested for security vulnerabilities by providing test data with known malicious properties to the application, and then observing the behavior of the application. For example, dynamic analysis may be used to test a web application that is accessed over a computer network, such as the Internet, by employing a “black-box tester” running on one computer that sends HTTP requests via a computer network to the web application that is hosted by another computer. The HTTP requests are configured with malicious test data payloads drawn from a library of test data payloads designed to test for security vulnerabilities such as SQL injection, cross-site scripting, and command injection.
A black-box tester could send all possible test data payloads to a computer software application to test for all possible security vulnerabilities. However, as the cost of sending many HTTP requests via a computer network is high in terms of time and bandwidth, commercial black-box testers that test web applications via computer networks typically send only a few dozen test payloads per HTTP parameter being tested in order to keep such costs down. Unfortunately, this often means that some security vulnerabilities go undetected.